HP Serviceguard Storage Management Suite Version A.02.01.01 for HP-UX 11i v3 Release Notes, Edition 1 (May 2010)

Use Case Examples
• Creating a mount point that is not nested:
cfsmntadm add cvm_dg1 lvol1 /cfs1 all=rw
• Creating a nested mount point:
cfsmntadm add cvm_dg1 lvol2 /cfs1/cfs2 all=rw (assuming cvm_dg1/lvol1
is already mounted on /cfs1, /cfs1/cfs2 becomes a nested mount point)
• Creating a ckpt nested mount point:
cfsmntadm add ckpt ckpt1 /cfs1 /cfs1/ckpt1 all=ro (if /cfs1 is a CFS mount
point, /cfs1/ckpt1 becomes a nested mount check point)
• Creating a snapshot nested mount point:
cfsmntadm add snapshot cvm_dg1 lvol3 /cfs1 /cfs1/snapshot1 snapshot1
all=ro (if /cfs1 is a CFS mount point, /cfs1/snapshot1 becomes a nested snapshot
mount point)
Removing a Node from a CFS Cluster
The following example illustrates how you would remove one node (node2) from a cluster
consisting of two CFS nodes (node1 and node2):
1. Disable AUTOSTART_CMCLD in /etc/rc.config.d/cmcluster on node2:
In /etc/rc.cofig.d/cmcluster, set AUTO_START_CMCLD=0
2. Halt node2:
cmhaltnode -f node2
3. Delete node2 from all of the application packages. Edit all application package ASCII files
and remove node2 from the package configuration. Re-apply the package configuration:
cmapplyconf –f –P app-pkg-ascii-file
4. Delete node2 from /mpt configuration – that is, modify the CFS mount point packages and
remove node2 from their configuration:
cfsmntadm modify /mpt delete node2
5. Delete node2 from dg configuration:
cfsdgadm modify dg delete node2
6. Delete node2 from the cluster. Edit the cluster ASCII file to remove node2 from the cluster
configuration:
cmapplyconf –f cluster1-ascii-file
Step 1 is a standard step that prevents CMCLD from starting upon reboot.
Step 4 must be repeated for all CFS mount-point MNP packages.
Step 5 must be repeated for all Disk Group MNP packages.
Step 6 results in a cluster consisting only of node 1 and all MNPs redefined on node 1.
SCSI-3 Persistent Reservations Support Not Required on Storage Devices
The Veritas Cluster Server (VCS) relies on support of SCSI-3 Persistent Reservations on storage
devices - this is part of the Veritas split brain prevention mechanism. HP Serviceguard uses a
different and more robust method to prevent split brain - storage devices do not need to support
SCSI-3 Persistent Reservations in an HP Serviceguard environment.
Product Issues and Limitations 33